From code to care: ANU research on AI is changing lives

Imagine you wake up hungry and have breakfast, then grab a snack later in the day. For most of us, it’s simple, spontaneous and done without much thought.

But for someone living with type 1 diabetes, every meal or snack means stopping to calculate carbs, measure, and inject insulin, usually about 20 minutes in advance.

It’s a daily reality Dr Chirath Hettiarachchi is determined to change.

“Our goal is to reduce that daily burden, or ideally, eliminate it altogether,” he says.

Working at the intersection of biomedical engineering and AI, Hettiarachchi, a Research Fellow with the School of Computing at The Australian National University (ANU), is developing cutting-edge automated insulin delivery systems.

“We use reinforcement learning – a type of AI where systems learn by trial and error – to build fully automated insulin delivery systems that continuously make insulin dosing decisions,” he explains.

For Hettiarachchi, whose background is in electronics, the path to health research may not be an immediately obvious one. But it was driven by a simple motivation: building technology that helps people.

“I’ve always enjoyed tackling research and engineering problems, even without much background knowledge,” he said.

“It’s incredibly rewarding to learn along the way and see people benefit from what I’ve built.”

After co-founding a healthcare startup developing predictive health tools,  Hettiarachchi took on further study exploring machine learning and biomedical signal processing for type 2 diabetes, before completing a PhD at ANU through the University’s Our Health in Our Hands (OHIOH)  initiative.

Now, the developed intelligent insulin delivery systems are being validated in pre-clinical trials using computer simulations, with close input from clinicians and patients along the way.

“We’re working hard to ensure the systems are safe, trustworthy and truly helpful in everyday settings,” he says.

That hard work is already earning recognition. Hettiarachchi was recently named a Highly Commended Outstanding New Researcher at the 2025 Canberra Health Annual Research Meeting (CHARM) Rising Star awards. He also received the 2025 ANU Early Career Research Impact Award, following international acclaim for his research, which won Best Scientific Paper at the 2024 Health Innovation Conference.

“Health research is inherently challenging, as developing and adopting new technologies often takes significant time. I’m motivated and content to be making my small contribution towards this goal.”

Beyond research, Hettiarachchi is passionate about education and accessibility. Under the OHIOH initiative, he developed a free, web-based educational tool known as CAPSML.

“We identified a gap between AI research and its translation,” he says. “CAPSML is both an education and research tool to bridge that divide.

“It enables clinicians and people with type 1 diabetes to explore the AI-driven systems we develop, without needing any programming background.”

Now used by more than 1,800 people across 55 countries, CAPSML has opened the door to student-led research too.

“I find it incredibly fulfilling to see how this project has grown,” he says. “So far, we’ve had one Master student and two Honours students at ANU, as well as four Honours students from overseas, all working on interesting new directions to improve our systems.

“I hope all our contributions will eventually feed into a system that eliminates patient burden and improves type 1 diabetes management.”

Hettiarachchi is a strong advocate for open science, sharing his team’s work online to improve accessibility and encourage progress.

“We’ve open-sourced all our code and tools so that others can continue our work,” he says.

“By sharing openly, we hope to accelerate research, foster growth across the field, and enable others to replicate and build on what we’ve done.”

Hettiarachchi says it has been an honour to collaborate as part of the ANU OHIOH team – an experience that’s reinforced the importance of designing with, not just for, communities.

“I’ve been trained to follow a problem-first approach, which I think is very important in applied research,” he says. “Technology-first approaches often lead to fancy yet impractical solutions, and they fail to recognise the expertise across disciplines.”

Looking ahead, Hettiarachchi is expanding his research into mental health, developing AI-supported treatments for depression.

“Health research is inherently challenging, as developing and adopting new technologies often takes significant time. I’m motivated and content to be making my small contribution towards this goal.”

For other early career researchers, particularly in health, his advice is simple: find your rhythm, take breaks and value your environment.

“Canberra is a calm and peaceful place for research. Especially for research like ours, which requires a lot of patience,” he says.

“It’s easy to feel worn down or challenged by how long things can take, but a nice walk outside always helps me to reset and stay motivated. Passion and patience go a long way – and a strong, supportive team makes all the difference in navigating your research.”